FAILED: patch "[PATCH] ext4: fix deadlock due to mbcache entry corruption" failed to apply to 5.15-stable tree

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Possible dependencies:

a44e84a9b776 ("ext4: fix deadlock due to mbcache entry corruption") 307af6c87937 ("mbcache: automatically delete entries from cache on freeing") 65f8b80053a1 ("ext4: fix race when reusing xattr blocks") fd48e9acdf26 ("ext4: unindent codeblock in ext4_xattr_block_set()") 6bc0d63dad7f ("ext4: remove EA inode entry from mbcache on inode eviction") 3dc96bba65f5 ("mbcache: add functions to delete entry if unused") 58318914186c ("mbcache: don't reclaim used entries") 4efd9f0d120c ("ext4: use kmemdup() to replace kmalloc + memcpy")

thanks,

greg k-h

------------------ original commit in Linus's tree ------------------

Quoted 1 line

From a44e84a9b7764c72896f7241a0ec9ac7e7ef38dd Mon Sep 17 00:00:00 2001

From: Jan Kara <[email protected]>
Date: Wed, 23 Nov 2022 20:39:50 +0100
Subject: [PATCH] ext4: fix deadlock due to mbcache entry corruption

When manipulating xattr blocks, we can deadlock infinitely looping inside ext4_xattr_block_set() where we constantly keep finding xattr block for reuse in mbcache but we are unable to reuse it because its reference count is too big. This happens because cache entry for the xattr block is marked as reusable (e_reusable set) although its reference count is too big. When this inconsistency happens, this inconsistent state is kept indefinitely and so ext4_xattr_block_set() keeps retrying indefinitely.

The inconsistent state is caused by non-atomic update of e_reusable bit. e_reusable is part of a bitfield and e_reusable update can race with update of e_referenced bit in the same bitfield resulting in loss of one of the updates. Fix the problem by using atomic bitops instead.

This bug has been around for many years, but it became *much* easier to hit after commit 65f8b80053a1 ("ext4: fix race when reusing xattr blocks").

Cc [email protected]

Fixes 6048c64b2609 ("mbcache: add reusable flag to cache entries")

Fixes 65f8b80053a1 ("ext4: fix race when reusing xattr blocks")

Reported-and-tested-by Jeremi Piotrowski <[email protected]>

Reported-by Thilo Fromm <[email protected]>

Link https://lore.kernel.org/r/[email protected]/ open here

Signed-off-by Jan Kara <[email protected]>

Reviewed-by Andreas Dilger <[email protected]>

Link https://lore.kernel.org/r/[email protected] open here

Signed-off-by Theodore Ts'o <[email protected]>

fs/ext4/xattr.c +2 -2

index 4d1c701f0eec..6bdd502527f8 100644
@@ -1281,7 +1281,7 @@ ext4_xattr_release_block(handle_t *handle, struct inode *inode,
 				ce = mb_cache_entry_get(ea_block_cache, hash,
 							bh->b_blocknr);
 				if (ce) {
-					ce->e_reusable = 1;
+					set_bit(MBE_REUSABLE_B, &ce->e_flags);
 					mb_cache_entry_put(ea_block_cache, ce);
 				}
 			}

@@ -2043,7 +2043,7 @@ ext4_xattr_block_set(handle_t *handle, struct inode *inode,
 				}
 				BHDR(new_bh)->h_refcount = cpu_to_le32(ref);
 				if (ref == EXT4_XATTR_REFCOUNT_MAX)
-					ce->e_reusable = 0;
+					clear_bit(MBE_REUSABLE_B, &ce->e_flags);
 				ea_bdebug(new_bh, "reusing; refcount now=%d",
 					  ref);
 				ext4_xattr_block_csum_set(inode, new_bh);

fs/mbcache.c +8 -6

index e272ad738faf..2a4b8b549e93 100644
@@ -100,8 +100,9 @@ int mb_cache_entry_create(struct mb_cache *cache, gfp_t mask, u32 key,
 	atomic_set(&entry->e_refcnt, 2);
 	entry->e_key = key;
 	entry->e_value = value;
-	entry->e_reusable = reusable;
-	entry->e_referenced = 0;
+	entry->e_flags = 0;
+	if (reusable)
+		set_bit(MBE_REUSABLE_B, &entry->e_flags);
 	head = mb_cache_entry_head(cache, key);
 	hlist_bl_lock(head);
 	hlist_bl_for_each_entry(dup, dup_node, head, e_hash_list) {

@@ -165,7 +166,8 @@ static struct mb_cache_entry *__entry_find(struct mb_cache *cache,
 	while (node) {
 		entry = hlist_bl_entry(node, struct mb_cache_entry,
 				       e_hash_list);
-		if (entry->e_key == key && entry->e_reusable &&
+		if (entry->e_key == key &&
+		    test_bit(MBE_REUSABLE_B, &entry->e_flags) &&
 		    atomic_inc_not_zero(&entry->e_refcnt))
 			goto out;
 		node = node->next;

@@ -284,7 +286,7 @@ EXPORT_SYMBOL(mb_cache_entry_delete_or_get);
 void mb_cache_entry_touch(struct mb_cache *cache,
 			  struct mb_cache_entry *entry)
 {
-	entry->e_referenced = 1;
+	set_bit(MBE_REFERENCED_B, &entry->e_flags);
 }
 EXPORT_SYMBOL(mb_cache_entry_touch);
 

@@ -309,9 +311,9 @@ static unsigned long mb_cache_shrink(struct mb_cache *cache,
 		entry = list_first_entry(&cache->c_list,
 					 struct mb_cache_entry, e_list);
 		/* Drop initial hash reference if there is no user */
-		if (entry->e_referenced ||
+		if (test_bit(MBE_REFERENCED_B, &entry->e_flags) ||
 		    atomic_cmpxchg(&entry->e_refcnt, 1, 0) != 1) {
-			entry->e_referenced = 0;
+			clear_bit(MBE_REFERENCED_B, &entry->e_flags);
 			list_move_tail(&entry->e_list, &cache->c_list);
 			continue;
 		}

include/linux/mbcache.h +7 -2

index 2da63fd7b98f..97e64184767d 100644
@@ -10,6 +10,12 @@
 
 struct mb_cache;
 
+/* Cache entry flags */
+enum {
+	MBE_REFERENCED_B = 0,
+	MBE_REUSABLE_B
+};
+
 struct mb_cache_entry {
 	/* List of entries in cache - protected by cache->c_list_lock */
 	struct list_head	e_list;

@@ -26,8 +32,7 @@ struct mb_cache_entry {
 	atomic_t		e_refcnt;
 	/* Key in hash - stable during lifetime of the entry */
 	u32			e_key;
-	u32			e_referenced:1;
-	u32			e_reusable:1;
+	unsigned long		e_flags;
 	/* User provided value - stable during lifetime of the entry */
 	u64			e_value;
 };

Quoted 8 lines · On Wed, Jan 04, 2023 at 04:07:31PM +0100, [email protected] wrote:

The patch below does not apply to the 5.15-stable tree.
If someone wants it applied there, or to any other stable or longterm tree, then please email the backport, including the original git commit id to <[email protected]>.

Possible dependencies:

Quoted 9 lines

a44e84a9b776 ("ext4: fix deadlock due to mbcache entry corruption") 307af6c87937 ("mbcache: automatically delete entries from cache on freeing") 65f8b80053a1 ("ext4: fix race when reusing xattr blocks") fd48e9acdf26 ("ext4: unindent codeblock in ext4_xattr_block_set()") 6bc0d63dad7f ("ext4: remove EA inode entry from mbcache on inode eviction") 3dc96bba65f5 ("mbcache: add functions to delete entry if unused") 58318914186c ("mbcache: don't reclaim used entries") 4efd9f0d120c ("ext4: use kmemdup() to replace kmalloc + memcpy")

Hi Jan,

What do you think of the backport I shared here:
https://lore.kernel.org/linux-ext4/20221122174807.GA9658@linuxonhyperv3.guj3yctzbm1etfxqx2vob5hsef.xx.internal.cloudapp.net/

Or do you think it makes more sense to backport some of the other patches listed above?